[Metrics][KV Offload] Add Prometheus metrics for SimpleCPUOffloadConnector (#57251)

Signed-off-by: Vincent <vincexxchan@gmail.com>
This commit is contained in:
mevince
2026-09-26 17:17:11 -07:00
committed by GitHub
parent 4be061c5af
commit cd1daf58a9
6 changed files with 903 additions and 6 deletions
+24
View File
@@ -45,6 +45,30 @@ The following metrics are exposed:
--8<-- "gen:metrics-nixl"
## Simple CPU Offload Connector Metrics
These metrics are exposed when the `SimpleCPUOffloadConnector` KV connector
is configured (e.g. `--kv-transfer-config='{"kv_connector":
"SimpleCPUOffloadConnector", "kv_role": "kv_both", "kv_connector_extra_config":
{"kv_offload_backend": "disk", "disk_path": "/mnt/nvme/kv"}}'`). They are
updated once per engine step.
Caveats to keep in mind when interpreting them:
- A "completed" store means the write syscalls returned; it is **not**
fsync-durable, and the disk backend's file is process-lifetime scratch,
unlinked at startup and shutdown.
- `save_outcomes_total` classifies eager-mode boundary hand-off stores;
lazy-mode stores are not classified.
- `used_blocks` counts blocks pinned by in-flight transfers or cache hits;
warm cached blocks that are evictable are not counted. Use the
`capacity_blocks` label of `simple_kv_offload_info` as the denominator.
- Counters and gauges are quantized to engine steps. They are reported by the
scheduler process and reflect engine-wide logical block counts, not values
pooled from individual tensor-parallel workers.
--8<-- "gen:metrics-simple-kv-offload"
## Model Flops Utilization (MFU) Performance Metrics
These metrics are available via `--enable-mfu-metrics`: